What’s unique about AccioIbis?

What is AccioIbis?

AccioIbis is an Artificial Intelligence-powered application that uses cutting edge NLP, Machine Learning and Neural Network algorithms to evaluate any writing and speaking in just seconds. AccioIbis is the premier platform to help students to prepare for various English language proficiency exams by evaluating their essay and speech quickly with high accuracy.

Here are the points which make AccioIbis unique –

Strict adherence to IELTS and TOEFL scoring criteria

The evaluation criteria and the scores displayed are all in compliance with the official criteria provided by the British Council and ETS.

Precision in scores

The scores obtained in the writing and speaking tasks on AccioIbis can be used as a benchmark for the expected scores in the final exam at IELTS or TOEFL. The scores that a student will obtain in the exam will be very close if not the same as the ones obtained on AccioIbis.

How are the scores evaluated?

The essay score is broken down into several categories such as lexical, cohesion, coherence, grammar, task achievement etc. With a detailed study of what each parameter comprises and how it can be calculated, a code was created to evaluate the score of each parameter for an essay.

The codes were run for thousands of essays and they were trained against the detailed scores of those essays. The grammar and spell check are done by 3 different codes that ensure that an outdated dictionary or a faulty grammar check doesn’t provide a bad score for a good essay. The three codes do multiple checks on the essay and hence make sure that Proper Nouns that are commonly used or widely known are not flagged as a spelling mistake. Apart from this, we also give a list of words or phrases that were repeated more than 3 times along with synonyms for the same as the repetition could reduce the lexical score.

Grammar and spelling mistakes are highlighted along with their replacements for the students’ reference. The Parts of Speech of the grammar mistakes are also identified and displayed so that students can identify their areas of weakness. NLP is applied extensively with the help of NLTK to identify coherence errors, sentence formation errors, subject-verb agreement errors, etc. apart from calculating the task achievement score. Apart from the detailed scores that are displayed, there are some other parameters that are also evaluated like the sentence and word score. These are hidden because they could potentially be manipulated to get a good overall score.

To summarize, the individual scores have their own breakdowns and a lot of parameters are calculated for each essay which is then fed to the machine learning model to get the overall score for each parameter. The scores of each parameter (displayed and hidden) are combined to decide the final score.

For speaking scores, individual parameters and overall scores are again as per British Council and ETS guidelines. The speaking scores are broken down into components like Fluency, Coherence, Lexical, Grammar and Pronunciation. Multiple Machine Learning models and Neural Network Models are used to achieve these scores.

With the craving for perfection becoming higher every day, AccioIbis ensures that the criteria, score breakdown and the scores displayed are accurate. We keep working to bring in new features to help more students practice and identify their strengths and weaknesses so that they can focus on improving their weaker areas to get a perfect score in the exam.